| Autor: |
Akhmedova, Shakhnaz, Stanovov, Vladimir, Ito, Kosuke, Mizuno, Yuna, Ohnishi, Fumiaki, Kamiya, Yukihiro |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 2023, Vol. 2700 Issue 1, p1-10, 10p |
| Abstrakt: |
This paper introduces a new neuro-fuzzy classifiers trained by recently developed modification of the Adaptive Gaining Sharing Knowledge (AGSK) optimization algorithm. Mentioned modification is based on applying the success-history based position adaption technique to improve the AGSK approach's ability to explore and exploit the search space. Proposed neuro-fuzzy classifiers were used to solve the opinion mining problems, to be more specific, they were applied to determine the customers' evaluation of coffee products based on various parameters. Besides, number of these parameters also varied with aim to define which of them are the most significant. It should be noted that these problems were distinguished due to different types of coffee beans used to prepare final products. Obtained results demonstrated the workability and usefulness of the neuro-fuzzy classifiers trained by the AGSK modification. Comparison to alternative approaches showed that the proposed technique outperforms them and, therefore, can be used instead of them for other classification problems. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
